Why I Am Not a Cognitive Psychologist
Author(s): B. F. Skinner
Source: Behaviorism, Vol. 5, No. 2 (Fall, 1977), pp. 1-10
Published by: Cambridge Center for Behavioral Studies
Stable URL: http://www.jstor.org/stable/27758892 .
Accessed: 21/10/2011 13:42
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Cambridge Center for Behavioral Studies is collaborating with JSTOR to digitize, preserve and extend access
to Behaviorism.
http://www.jstor.org
Why I am Not a Cognitive Psychologist
B.F. Skinner
Harvard University
The variables of which human behavior is a function lie in the environment. We
distinguish between (1) the selective action of that environment during the evolution of
the species, (2) its effect in shaping and maintaining the repertoire of behavior which
converts each member of the species into a person, and (3) its role as the occasion upon
which behavior occurs. Cognitive psychologists study these relations between organism
and environment, but they seldom deal with them directly. Instead they invent internal
surrogates which become the subject matter of their science.
Take, for example, the so-called process of association. In Pavlov's experiment a
hungry dog hears a bell and is then fed. If this happens many times, the dog begins to
salivate when it hears the bell. The standard mentalistic explanation is that the dog
"associates" the bell with the food. But itwas Pavlov who associated them! "As
sociate" means to join or unite. The dog merely begins to salivate upon hearing the bell.
We have no evidence that itdoes so because of an internal surrogate of the contingencies.
In the "association of ideas" the ideas are internal replicas of stimuli towhich I shall
return. Ifwe have eaten lemons, we may taste lemon upon seeing a lemon or see a lemon
upon tasting lemon juice, but we do not do this because we associate the flavor with the
appearance. They are associated in the lemon. "Word associations" are at least correctly
named. Ifwe say "home" when someone says "house," it is not because we associate
the two words but because they are associated in daily English usage. Cognitive
association is an invention. Even if itwere real, itwould go no further toward an
explanation than the external contingencies upon which it ismodelled.
Another example is abstraction. Consider a simple experiment. A hungry pigeon can
peck any one of a number of panels bearing the names of colors?"white," "red,"
"blue,'' and so on, and thepecks are reinforced with small amounts of food. Any one of a
number of objects?blocks, books, flowers, toy animals, and so on?can be seen in an
adjacent space. The following contingencies are then arranged: whenever the object is
white, no matter what its shape or size, pecking only the panel marked "white" is
reinforced; whenever the object is red, pecking only the panel marked "red" is rein
forced; and so on. Under these conditions the pigeon eventually pecks the panel marked
"white" when the object iswhite, the panel marked "red" when the object is red, and so
on. Children are taught to name colors with similar contingencies, and we all possess
comparable repertoires sustained by the reinforcing practices of our verbal environ
ments.
But what is said tobe going on in themind? Karl Popper (1957) has put a classical issue
thisway: "We can say either that (1) the universal term "white" is a label attached to a set
1
B. F. Skinner
of things, or that (2) we collect the set because they share an intrinsic property of
"whiteness." Popper says the distinction is important; natural scientists may take the
firstposition but social sciencists must take the second. Must we say, then, that thepigeon
is either attaching a universal term to a set of things or collecting a set of things because
they share an intrinsic property? Clearly, it is the experimenter not the pigeon who
"attaches" the white key to the white objects displayed and who collects the set of
objects on which a single reinforcing event ismade contingent. Should we not simply
attribute the behavior to the experimental contingencies? And ifso, why not for children
or ourselves? Behavior comes under the control of stimuli under certain contingencies of
reinforcement. Special contingencies maintained by verbal communities produce
"abstractions." We do attach physical labels to physical things and we collect physical
objects according to labels properties, but comparable cognitive processes are inven
tions, which even if real, would be no closer to an explanation than the external
contingencies.
Another cognitive account of the same data would assert thata person, ifnot a pigeon,
forms an abstract idea or develops a concept of color. The development of concepts is an
especially popular cognitive field. (The horticultural metaphor minimizes contributions
from the environment. We may hasten the growth of the mind but we are no more
responsible for its final character than farmers for the character of the fruits and
vegetables they so carefully nourish.) Color vision is part of the genetic endowment of
most people, and itdevelops or grows in a physiological sense, possibly to some extent
after birth, Nevertheless, most stimuli acquire control because of their place in con
tingencies of reinforcement. As the contingencies become more complex, they shape and
maintain more complex behavior. It is the envirionment that develops, not a mental or
cognitive possession.
A passage from a recent discussion of the development of sexual identity in a child
might be translated as follows: "The child forms a concept based upon what it has
observed and been told of what itmeans to be a boy or girl." (A child's behavior is
affected by what ithas observed and been told about being a boy or girl.)' 'This concept is
oversimplified, exaggerated, and stereotyped." (The contingencies affecting the be
havior are simplified and exaggerated and involve stereotyped behavior on the part of
parents and others.) "As the child develops cognitively, itsconcepts, and consequently
its activities, become more sophisticated and realistic." (As the child grows older, the
contingencies become more subtle and more closely related to the actual sex of the child.)
Children do not go around forming concepts of their sexual identity and "consequently"
behaving in special ways; they slowly change theirbehavior as people change theways in
which they treat them because of their sex. Behavior changes because the contingencies
change, not because a mental entity called a concept develops.
Many mentalistic or cognitive terms refer not only to contingencies but to the
behavior they generate. Terms like "mind," "will," and "thought" are often simply
synonyms of "behavior." An historian writes: "what may be called a stagnation of
thought prevailed, as though themind, exhausted after building up the spiritual fabric of
theMiddle Ages, had sunk into inertia." Exhaustion is a plausible metaphor when a
2
Why I AmNot a CognitivePsychologist
quiet period follows an active one, but itwas behavior that became stagnant and inert,
presumably because the contingencies changed. Certain social conditions ("the spiritual
fabric of theMiddle Ages") made people active. A second set of conditions, possibly
produced by the very behavior generated by the first,made them much less so. To
understand what actually happened, we should have to discover why the contingencies
changed, not why thought became stagnant or inert.
Behavior is internalized as mental life when it is too slight to be observed by
others?when, as we say, it is covert. A writer has pointed out that "the conductor of an
orchestra maintains a certain even beat according to an internal rhythm, and he can divide
that beat in half again and again with an accuracy rivaling any mechanical instrument."
But is there an internal rhythm? Beating time isbehavior. Parts of the body often serve as
pendulums useful indetermining speed, as when the amateur musician beats time with a
foot or the rock player with the whole body, but other well-timed behavior must be
learned. The conductor beats time steadily because he has learned to do so under rather
exacting contingencies of reinforcement. The behavior may be reduced in scale until it is
no longer visible to others. It is still sensed by the conductor, but it is a sense of behavior
not of time. The history of "man's development of a sense of time" over the centuries is
not a matter of cognitive growth but of the invention of clocks, calendars, and ways of
keeping records?in other words, of an environment that "keeps time."
When an historian reports that in a given period "a wealthy, brilliant, and traditional
governing class lost itswill,'' he is reporting simply that itstopped acting like a wealthy,
brilliant, and traditional governing class. Deeper changes are suggested by the term
"will" but they are not identified. They could not have been changes in particular
people, since the period lasted more than one lifetime. What changed were presumably
the conditions affecting the behavior of members of the class. Perhaps they lost their
money; perhaps competing classes became more powerful.
Feelings, or the bodily conditions we feel, are commonly taken as the causes of
behavior. We go for a walk "because we feel like going." It is surprising how often the
futility of such an explanation is recognized. A distinguished biologist, C. H. Wad
dington (1974), reviewing a book by Tinbergen, writes as follows:
It is not clear how far he (Tinbergen) would go along with the argument of
one of themost perceptive critical discussions of ethology by Suzanne Langer,
who argues thateach step ina complex structure of behavior iscontrolled, not by
a hierarchical set of neural centers, but by the immediate feelings of the animal.
The animal, she claims, does the next thing in the sequence, not to bring about a
useful goal, or even as amove toward an enjoyable consummation, but because it
actually feels like doing itat themoment.
Evidently Waddington himself goes along part way with this "perceptive view."
But suppose Langer is right. Suppose animals simply do what they feel like doing?
What is the next step inexplaining their behavior? Clearly, a science of animal behavior
must be replaced or supplemented by a science of animal feelings. It would be as
extensive as the science of behavior because therewould presumably be a feeling for each
act. But feelings are harder to identify and describe than the behavior attributed to them,
and we should have abandoned an objective subject matter in favor of one of dubious
status, accessible only through necessarily defective channels of introspection. The
3
B,F. Skinner
contingencies would be the same. The feelings and the behavior would have the same
causes.
A British statesman recently asserted that the key to crime in the streets was
"frustration." Young people mug and rob because they feel frustrated. But why do they
feel frustrated? One reason may be thatmany of them are unemployed, either because
they do not have the education needed to get jobs or because jobs are not available. To
solve the problem of street crime, therefore, we must change the schools and the
economy. But what role is played in all this by frustration? Is it the case thatwhen one
cannot get a job one feels frustrated and thatwhen one feels frustrated one mugs and robs,
or is it simply the case thatwhen one cannot earn money, one ismore likely to steal
it?and possibly to experience a bodily condition called frustration?
Since many of the events which must be taken into account inexplaining behavior are
associated with bodily states that can be felt, what is felt may serve as a clue to the
contingencies. But the feelings are not the contingencies and cannot replace them as
causes.
By itsvery nature operant behavior encourages the invention of mental or cognitive
processes said to initiate action. In a reflex, conditioned or unconditioned, there is
conspicuous prior cause. Something triggers the response. But behavior that has been
positively reinforced occurs upon occasions which, though predisposing, are never
compelling. The behavior seems to start up suddenly, without advance notice, as if
spontaneously generated. Hence the invention of such cognitive entities as intention,
purpose, or will. The same issues were debated with respect to the theory of evolution and
for the same reason: selection is a special causal mode not easily observed. Because
controlling circumstances which lie in an organism's history of reinforcement are
obscure, themental surrogate gets itschance. Under positive reinforcement we do, as we
say,what we are free todo; hence the notion of freewill as an initiating condition. (I think
itwas Jonathan Edwards who said thatwe believe infreewill because we know about our
behavior but not about its causes.)
When we do not know why people do one thing rather than another, we say that they
"choose" or "make decisions." Choosing originally meant examining, scrutinizing, or
testing. Etymologically, deciding means cutting off other possibilities, moving in a
direction from which there is no return. Choosing and deciding are thus conspicuous
forms of behavior, but cognitive psychologists have nevertheless invented internal
surrogates. Anatole Rapaport (1973) puts it this way: "A subject in a psychological
experiment is offered a choice among alternatives and selects one alternative over
others." When this happens, he says, "common sense suggests that he is guided by a
preference." Common sense does indeed suggest it,and so do cognitive psychologists,
but where and what is a preference? Is itanything more than a tendency to do one thing
rather than another? When we cannot tellwhence thewind cometh and wither itgoeth,
we say that it"bloweth where it listeth," and common sense, ifnot cognitive psychology,
thus credits itwith a preference. (List, by the way, is an example of a term with a
physical referent used to refer to a mental process. Itmeans, of course, to lean?as in the
listof a ship. And since things usually fall in the direction inwhich they are leaning, we
4
Why I AmNot a CognitivePsychologist
say that people lean toward a candidate in an election as a rough way of predicting how
theywill vote. The same metaphor is found in"inclincation"; we are "inclined" to vote
forX. But itdoes not follow thatwe have internal leanings and inclinations which affect
our behavior.)
"Intention" is a rather similar term which once meant stretching. The cognitive
version is a critical issue incurrent linguistics. Must the intention of the speaker be taken
into account? In an operant analysis verbal behavior is determined by the consequences
which follow in a given verbal environment, and consequences are what cognitive
psychologists are really talking about when they speak of intentions. All operant
behavior "stretches forward" a future even though the only consequences responsible
for its strength have already occurred. I go to a drinking fountain "with the intention of
getting a drink ofwater'' in the sense that Igo because in thepast I have got a drink when I
have done so. (I may go for the first time, following directions, but that is not an
exception; it is an example of rule-governed behavior, of which more later.)
So much for the cognitive internalization of contingencies of reinforcement and the
invention of cognitive causes of behavior. Far more damaging to an effective analysis is
the internalization of the environment. The Greeks invented themind to explain how the
real world could be known. For them, toknow meant tobe acquainted with, tobe intimate
with. The term cognition itself is related to coitus, as in theBiblical sense inwhich aman
is said to know a woman. Having no adequate physics of light and sound nor any
chemistry of taste and odor, theGreeks could not understand how a world outside the
body, possibly some distance away, could be known. There must be internal copies.
Hence cognitive surrogates of the'real world.
The distinction between reality and conscious experience has been made so often that
itnow seems self-evident. Fred Attneave (1974) has recently written that "the statement
that theworld as we know it isa representation is, I think, a truism?there is really no way
inwhich itcan be wrong." But there are at least twoways, depending upon themeaning.
If the statement means thatwe can know only representations of the outside world, it is a
"truism" only ifwe are not our bodies but inhabitants located somewhere inside. Our
bodies are incontact with the real world and can respond to itdirectly, but ifwe are tucked
away up in the head, we must be content with representations.
Another possible meaning is thatknowing is the very process of constructing mental
copies of real things, but if that is the case how do we know the copies? Do we make
copies of them! And is that regress infinite?
Some cognitive psychologists recognize that knowing is action but try tomake the
point by appealing to another mental surrogate. Knowledge is said to be "a system of
propositions." According toone writer, "when we use theword 'see' we refer to a bridge
between a pattern of sensory stimulation and knowledge which is propositional." But
"propositional" is simply a laundered verion of "behavioral," and the "bridge" is
between stimuli and behavior and was built when the stimuli were part of the contingen
cies.
Representational theories of knowledge are modeled on practical behavior. We do
make copies of things.We construct representational works of art, because looking at
5
B. F. Skinner
them is reinforced inmuch the same way as looking at what they represent. We make
maps, because our behavior in following them is reinforced when we arrive at our
destination in the mapped territory. But are there internal surrogates? When we day
dream, do we first construct copies of reinforcing episodes which we thenwatch, or do
we simply see things once again? And when we learn to get about ina given territory,do
we construct cognitive maps which we then follow or do we follow the territory? Ifwe
follow a cognitive map, must we learn to do so, and will that require a map of themap?
There is no evidence of themental construction of images to be looked at or maps to be
followed. The body responds to theworld, at the point of contact; making copies would
be a waste of time.
Knowledge isa key term incognitive theory, and itcovers a good deal of ground. It is
often contrasted with perception. We are said tobe able to see that there are three dots on
a card but only to know that there are thirteen after counting them, even though counting
is a form of behavior. After noting that one spiral can be seen to be continuous but that
another can be discovered to be so.only by tracing, Bela Julesz (1975) has said that"any
visual task thatcannot be performed spontaneously, without effortor deliberation, can be
regarded as a cognitive task rather than as a perceptual one," though all the steps in that
example are also clearly behavioral.
44
Knowing how todo something" isan internal surrogate of behavior in itsrelation to
contingencies. A child learns to ride a bicycle and is then said to possess knowledge of
how to ride. The child's behavior has been changed by the contingencies of reinforce
ment maintained by bicycles; the child has not taken possession of the contingencies.
To speak of knowing about things is also to construct an internal surrogate of
contingencies. We watch a football game and are then said topossess knowledge of what
happened. We read a book and are said to know what it is about. The game and the book
are somehow
4
4represented" inour minds: we are "in possession of certain facts." But
the evidence is simply thatwe can describe what happened at the game and reportwhat
the book was about. Our behavior has been changed, but there isno evidence thatwe have
acquired knowledge. To be "in possession of the facts" isnot to contain the facts within
ourselves but to have been affected by them.
Possession of knowledge implies storage, a field inwhich cognitive psychologists
have constructed a great many mental surrogates of behavior. The organism is said to take
in and store the environment, possibly in some processed form. Let us suppose that a
young girl saw a picture yesterday and when asked todescribe ittoday, does so.What has
happened? A traditional answer would run something like this:when she saw the picture
yesterday the girl formed a copy inhermind (which, infact,was really all she saw). She
encoded it ina suitable form and stored it inhermemory, where itremained until today.
When asked to describe the picture today, she searched her memory, retrieved the
encoded copy, and converted it into something like the original picture, which she then
looked at and described. The account ismodelled on the physical storage ofmemoranda.
We make copies and other records, and respond to them. But do we do anything of the
sort inour minds?
6
Why I Am Not a Cognitive Psychologist
Ifanything is"stored," it isbehavior. We speak of the "acquisition" of behavior, but
inwhat form is itpossessed? Where is behavior when an organism is not behaving?
Where at the present moment, and inwhat form, is the behavior I exhibit when I am
listening tomusic, eating my dinner, talking with a friend, taking an early morning walk,
or scratching an itch?A cognitive psychologist has said that verbal behavior is stored as
"lexical memories." Verbal behavior often leaves public records which can be stored in
files and libraries, and themetaphor of the storage is therefore particularly plausible. But
is the expression any more helpful than saying thatmy behavior in eating my dinner is
stored as prandial memories, or scratching an itch as a prurient memory? The observed
facts are simple enough: I have acquired a repertoire of behavior, parts of which Idisplay
upon appropriate occasions. The metaphor of storage and retrieval goes well beyond
those facts.
The computer, together with information theory as designed to deal with physical
systems, has made the metaphor of input-storage-retrieval-output fashionable. The
struggle tomake machines that think like people has had the effect of supporting theories
inwhich people think likemachines. Mind has recently been defined as "the system of
organizations and structures ascribed to an individual that processess inputs
... and
provides output to the various subsystems and the world." But organizations and
structures of what? (The metaphor gains power from the way inwhich itdisposes of
troublesome problems. By speaking of input one can forget all the travail of sensory
psychology and physiology; by speaking of output one can forget all the problems of
reporting and analyzing action; and by speaking of the storage and retrieval of informa
tion one can avoid all the difficult problems of how organisms are indeed changed by
contact with their environments and how those changes survive.)
Sensory data are often said to be stored as images, much like the images said to
represent the real world. Once inside, they are moved about for cognitive purposes.
There is a familiar experiment on color generalization inwhich a pigeon pecks at a disk
of, say, green light, the behavior being reinforced on a variable interval schedule. When a
stable rate of responding develops, no further reinforcements are given, and the color of
the disk is changed. The pigeon responds to another color at a ratewhich depends upon
how much itdiffers from the original; rather similar colors evoke fairly high rates, very
different colors low rates. A cognitive psychologist might explain thematter in thisway:
The pigeon takes in a new color (as "input"), retrieves the original color frommemory,
where ithas been stored in some processed form, puts the two colored images side by side
so that theymay be easily compared, and after evaluating the difference, responds at the
appropriate rate. But what advantage isgained by moving from a pigeon that responds to
different colors on a disk to an inner pigeon that responds to colored images in itsmind?
The simple fact is that because of a known history of reinforcement, different colors
control different rates.
The cognitive metaphor isbased upon behavior in the real world. We store samples of
material and retrieve and compare them with other samples. We compare them in the
literal sense of putting them side by side tomake differences more obvious. And we
respond to different things in different ways. But that is all. The whole field of the
processing of information can be reformulated as changes in the control exerted by
stimuli.
7
B. F. Skinner
The storage of factual knowledge raises another problem. When I learn, say, to take
apart the rings of a puzzle, itseems unlikely that I storemy knowledge of how todo so as a
copy of thepuzzle or of the contingencies thepuzzle maintains for those trying to solve it.
Instead cognitive theory holds that I store a rule. Rules are widely used as mental
surrogates of behavior, inpart because they can be memorized and hence "possessed,"
but there is an important difference between rules and the contingencies they describe.
Rules can be internalized in the sense thatwe can say them to ourselves, but indoing so
we do not internalize the contingencies.
Imay learn to solve the puzzle ineither of twoways. Imay move the rings about until
I hit upon a response that separates them. The behavior will be strengthened, and if I do
the same thing a number of times, Iwill eventually be able to take the rings apart quickly.
My behavior has been shaped and maintained by its effects on the rings. Imay, on the
other hand, simply follow printed directions supplied with the puzzle. The directions
describe behavior that separates the rings, and if I have already learned to follow
directions, I can avoid the possibly long process of having my behavior shaped by the
contingencies.
Directions are rules. Like advice, warnings, maxims, proverbs, and governmental
and scientific laws, they are extremely important parts of a culture, enabling people to
profit from the experience of others. Those who have acquired behavior through expo
sure to contingencies describe the contingencies, and others then circumvent exposure by
behaving in theways described. But cognitive psychologists contend that something of
the same sort happens internallywhen people learn directly from the contingencies. They
are said to discover rules which they themselves then follow. But rules are not in the
contingencies, normust they be "known'' by those who acquire behavior under exposure
to them. (We are lucky that this could be so, since rules are verbal products which arose
very late in the evolution of the species.)
The distinction between rules and contingencies is currently important in the field of
verbal behavior. Children learn to speak through contact with verbal communities,
possibly without instruction. Some verbal responses are effective and others not, and
over a period of time more and more effective behavior is shaped and maintained. The
contingencies having this effect can be analyzed. A verbal response "means" something
in the sense that the speaker is under the control of particular circumstances; a verbal
stimulus "means" something in the sense that the listener responds to it in particular
ways. The verbal community maintains contingencies of such a nature that responses
made upon particular occasions serve as useful stimuli to listeners who then behave
appropriately to the occasions.
More complex relations among the behaviors of speaker and listener fall within the
fields of syntax and grammar. Until the time of theGreeks, no one seems to have known
that therewere rules of grammar, although people spoke grammatically in the sense that
they behaved effectively under the contingencies maintained by verbal communities, as
children today learn to talk without being given rules to follow. But cognitive
psychologists insist that speakers and listeners must discover rules for themselves. One
authority, indeed, has defined speaking as "engaging in a rule-governed form of inten
tional behavior." But there is no evidence that rules play any part in the behavior of the
ordinary speaker. By using a dictionary and a grammar we may compose acceptable
8
Why I AmNot a CognitivePsychologist
sentences in a language we do not otherwise speak, and we may occasionally consult a
dictionary or a grammar in speaking our own language, but even so we seldom speak by
applying rules. We speak because our behavior is shaped and maintained by the practices
of a verbal community.
Having moved the environment inside the head in the form of conscious experience
and behavior in the form of intention, will, and choice, and having stored the effects of
contingencies of reinforcement as knowledge and rules, cognitive psychologists put them
all together to compose an internal simulacrum of the organism, a kind of Doppleganger,
not unlike the classical homunculus, whose behavior is the subject of what Piaget and
others have called "subjective behaviorism." The mental apparatus studied by cognitive
psychology is simply a rather crude version of contingencies of reinforcement and their
effects.
Every so-called cognitive process has a physical model. We associate things by
putting them together.We store memoranda and retrieve them for later use. We compare
things by putting them side by side to emphasize differences. We discriminate things one
from another by separating them and treating them indifferent ways. We identify objects
by isolating them from confusing surroundings. We abstract sets of items from complex
arrays. We describe contingencies of reinforcement inrules. These are the actions of real
persons. It is only in the fanciful world of an inner person that they become mental
processes.
The very speed with which cognitive processes are invented to explain behavior
should arouse our suspicions. Moliere made a joke of amedical example more than three
hundred years ago: "I am asked by the learned doctors for the cause and reason why
opium puts one to sleep, towhich I reply that there is in ita soporific virtue the nature of
which is to lull the senses." Moliere's candidate could have cited evidence from
introspection, invoking a collateral effect of the drug, by saying: "To which I reply that
opium makes one feel sleepy.'' But the soporific virtue itself is a sheer invention, and it is
not without current parallels.
A conference was recently held inEurope on the subject of scientific creativity. A
report published inScience (1974) begins by pointing out thatmore than ninety percent of
scientific innovation has been accomplished by fewer than ten percent of all scientists.
The next sentence might be paraphrased in thisway: "I am asked by the learned doctors
for the cause and reason why this should be so, towhich I reply that it is because only a
few scientists possess creativity." Similarly, "I am asked by the learned doctors for the
cause and reason why children learn to talk with great speed, towhich I reply that it is
because they possess linguistic competence." Moliere's audiences laughed.
Cognitive psychologists have two answers to the charge that themental apparatus is
ametaphor or construct. One is thatcognitive processes are known through introspection.
Do not all thinking persons know that they think?And ifbehaviorists say they do not, are
they not either confessing a low order of mentality or acting inbad faith for the sake of
their position? No one doubts that behavior involves internal processes; the question is
9
B. F. Skinner
how well they can be known through introspection. As I have argued elsewhere,
self-knowledge, consciousness, or awareness became possible only when the species
acquired verbal behavior, and thatwas very late in itshistory. The only nervous systems
then available had evolved for other purposes and did not make contact with themore
important physiological activities. Those who see themselves thinking see littlemore
than theirperceptual and motor behavior, overt and covert. They could be said toobserve
the results of "cognitive processes'' but not the processes themselves?a "stream of
consciousness" but not what causes the streaming, the "image of a lemon" but not the
act of associating appearance with flavor, their use of an abstract termbut not the process
of abstraction, a name recalled but not its retrieval frommemory, and so on. We do not,
through introspection, observe the physiological processes through which behavior is
shaped and maintained by contingencies of reinforcement.
But physiologists observe them and cognitive psychologists point to resemblances
which suggest that they and the physiologists are talking about the same things. The very
fact thatcognitive processes are going on inside the organism suggests that the cognitive
account is closer to physiology than the contingencies of reinforcement studied by those
who analyze behavior. But ifcognitive processes are simply modelled upon the environ
mental contingencies, the fact that they are assigned to space inside the skin does not
bring them closer to a physiological account. On the contrary, the fascination with an
imagined inner life has led to a neglect of the observed facts. The cognitive constructs
give physiologists a misleading account of what theywill find inside.
In summary, then, I am not a cognitive psychologist for several reasons. I see no
evidence of an innerworld of mental life relative either to an analysis of behavior as a
function of environmental forces or to the physiology of the nervous system. The
respective sciences of behavior and physiology will move forward most rapidly if their
domains are correctly defined and analyzed.
I am equally concerned with practical consequences. The appeal to cognitive states
and processes is a diversion which could well be responsible formuch of our failure to
solve our problems. We need to change our behavior and we can do so only by changing
our physical and social environments. We choose thewrong path at the very startwhen
we suppose thatour goal is to change the "minds and hearts of men and women" rather
than theworld inwhich they live.
REFERENCES
Attneave, Fred, American Psychologist, July, 1974.
Julesz, Bela, Scientific American, April, 1975.
Popper, Karl, Poverty of Historicism, London, 1957.
Rapaport, Anatol, Experimental Games and their Uses in Psychology, General Learning Press, 1973,
Science, June 21, 1974.
Waddington, C. H., New York Review, February 3, 1974.
10